BIDS Projekt

Gruppe 3

Constanze Leichtweiß
Ivanna Marchesini
Christoph Kovacs
Enrico Wondra

Intro
[Ivanna, Enrico]

Überblick [Ivanna]

Wirkungsweise [Enrico]

flowchart TD
    RTK["Rezeptor-Tyrosinkinase (zB EGFR, VEGFR)"]
    RAS["RAS (GTPase)"]
    RAF["RAF-Kinase (inkl. BRAF)"]
    MEK["MEK1/2-Kinase"]
    ERK["ERK1/2-Kinase"]
    Prolif["Zellproliferation und Überleben"]

    RTK --> RAS --> RAF --> MEK --> ERK --> Prolif

    SORA["Sorafenib (RAF-Hemmer)"] -.-> RAF
    TRAM["Trametinib (MEK-Hemmer)"] -.-> MEK

    style SORA fill:#fdd835,stroke:#f57f17,stroke-width:2px
    style TRAM fill:#81d4fa,stroke:#0288d1,stroke-width:2px

DGE:
Differentielle Gen-Expressions Analyse
[Constanze]

Set 1

Set 2

Group Separability [Chris]

Optimale Anzahl Komponenten für PCA

  • Nur Top 500 Gene miteinbezogen
  • 2-3 Komponenten optimal

  • Klare Separierbarkeit entlang PC1
  • Stabile Kontrollgruppe
  • Sora deutlich von DMSO unterschieden
  • Tram mit In-Sample Unterschieden (PC2)
  • Problem: Misst nur linearen Zusammenhang
  • PC3 erklärt nur knapp 7% der Gesamtvarianz
  • Keine bessere Separabilität

t-distributed Stochastic Neighbor Embedding

  • Misst lokalen nicht-linearen Zusammenhang
  • Deutliche Cluster der drei Treatments
  • Trametinib kohärenter als bei PCA

Uniform Manifold Approximation and Projection

  • Misst lokalen und globalen Zusammenhang
  • “Gradientenstruktur” DMSO → Tram → Sora
  • Problem: Zu wenige Samples

GSE: Gene-set Enrichment Analysis
[Chris, Constanze]

Übersicht [Chris]

  • Sorafenib ca. 3500 hoch- und runterregulierte Gene
  • Trametinib nur ca. 800

Hochregulierte Gene [Constanze]

  • Up-Regulated genes Plot (Sora, Tram, Together?) gibt es überhaupt Tram upregulated?

Runterregulierte Gene [Constanze]

  • Down-Regulated genes Plot (Sora, Tram, Together?)

Gene Ontology (GO) and KEGG Pathway analysis
[Ivanna, Enrico]

Hallmark

Summary [Ivanna]

  • Variant Calling